Weak and Strong Learning of Context-Free Grammars - 2012
Explore a comprehensive lecture on the advancements in unsupervised learning of context-free grammars using distributional techniques. Delve into the challenges of moving from weak learning algorithms to strong algorithms capable of learning correct structures. Examine a new theoretical approach based on transformations of grammars through morphisms of algebraic structures. Discover how this model leads to predictions about syntactic structure in natural languages. Learn about the syntactic concept lattice and its role as a basis for weak learning algorithms for CFGs. Gain insights from Alexander Clark, a researcher in grammatical inference, theoretical and mathematical linguistics, and unsupervised learning from the Department of Computer Science at Royal Holloway, University of London.